How long can a sourdough starter go without feeding? (And how to revive a neglected one)

How long a starter lasts, by storage method

A sourdough starter left in the refrigerator will still be alive after several months and will not necessarily be able to raise bread. Those are separate questions and almost nobody separates them.

The named authorities give strikingly different answers, and the peer-reviewed data lands squarely on the conservative end.

The convergence is the interesting part. Maurizio Leo's three weeks, Tom Cucuzza's thirty days and the laboratory's thirty days all land in the same place, arrived at independently. King Arthur's "months" is not wrong either; it is answering a different question.

At room temperature the timeline compresses hard. King Arthur's own staged description has hooch appearing at about a week, darker hooch at about a month, a dark grey corrugated skin after several months, and something closer to a chalk-like block after a year or more. A refrigerated starter, meanwhile, can go one to two weeks between feedings as ordinary maintenance rather than as neglect.

Neglect costs you rise before it costs you sourness

This is the finding that reframes everything else on the page, and we have not seen it stated anywhere in popular sourdough writing.

Lattanzi, Minervini and Gobbetti (2014) compared refrigerated, frozen and dried storage of a type I wheat sourdough and then reactivated each one. Refrigeration gave satisfactory results until 30 days. After 60 to 90 days, no storage condition produced a reactivated sourdough with leavening capacity comparable to unstored sourdough. But, and this is the part that matters, all of the sourdoughs maintained their acidification capacity when reactivated.

Souring survives. Rising does not. A starter that has been neglected for two or three months can come back smelling exactly right, tasting exactly right, dropping pH on schedule, and still fail to inflate a loaf.

That single asymmetry explains the most common frustration people report after a revival: the starter looks and smells healthy after a few feeds, the bread comes out flat, and nothing in the usual troubleshooting advice accounts for it. The smell is coming from the bacteria, which came through the neglect in better shape than the yeast did.

The same pattern shows up in every stress condition anyone has measured. Drying, freezing and acid all take the yeast out faster than the bacteria. If you want the mechanism rather than the observation, the next section has it.

What accumulated acid actually does to a starter

The acid a starving starter accumulates stops the culture growing long before it kills anything, and it stops the yeast at a lower concentration than it stops the bacteria.

Gänzle, Ehmann and Hammes (1998) measured the tolerances directly. Lactobacillus sanfranciscensis tolerated more than 160 mmol per liter of undissociated acetic acid and grew across a pH range of 3.9 to 6.7. Candida milleri, the yeast, was completely inhibited at 150 mmol per liter, while being essentially indifferent to pH between 3.5 and 7.

So as an unfed starter accumulates acetic acid, the yeast hits its wall first and the bacteria keep going. The culture does not fail evenly. It shifts.

It also explains why bigger feeds beat more frequent small ones during revival. A 1:5:5 feed drops the concentration of accumulated acid immediately, while twice-daily 1:1:1 keeps adding food to a jar that is still chemically hostile. The same logic drives the fix for a starter that smells like acetone.

The revival protocol, by how long it has been neglected

Match the protocol to the neglect. A one-week starter needs one feed; a six-month starter needs a week of patience and a decision point at the end of it.

The authorities agree on the direction and disagree on how aggressive to be. Cucuzza discards roughly 80 percent at each peak across three to five cycles. King Arthur's protocol is a plain 50 grams of starter, 50 grams of flour and 50 grams of water, left 12 hours, repeated. The Perfect Loaf asks for at least two and preferably three regular feedings. All three are dilution schedules; they differ only in slope.

Two pieces of expectation-setting that matter more than the ratios. King Arthur's documented case of a starter unfed for a full month at room temperature showed no activity at all after the first feeding, then nearly doubled within five hours after the second. And the volunteers who distribute Carl Griffith's heritage starter tell recipients plainly that old starters can take a week of repeated feeding to revive. A single failed feed is not a verdict.

Judge recovery by trend rather than by an absolute threshold. King Arthur's benchmark of doubling within 6 to 8 hours is the right target, and a starter three days into a long revival will fail it while recovering perfectly well. What you want to see is each feed peaking sooner than the last.

Freezing does not kill a sourdough starter

Freezing costs a starter a large fraction of its viable cells and does not kill the culture. One widely repeated claim to the contrary is not supported by the published comparison.

King Arthur states flatly that freezing will kill your starter. In Lattanzi's comparison of storage methods, freezing at minus 20°C was the method that performed best at 90 days, with bread made from the frozen sourdough approaching that made from unstored sourdough, and it enabled partial preservation of the starter lactic acid bacteria out to 90 days.

The cost is real. A 2025 review in Foods reports freezing reducing LAB and yeast viable cells by approximately 4 logarithmic values, and notes that freezing at minus 20°C is harder on cells than minus 80°C because larger ice crystals form at the higher temperature.

Losing four logs is a serious hit, and a frozen starter will need several feeds to rebuild. But four logs down from roughly 9 log CFU per gram is not zero, and the loaf evidence says it comes back. We would not choose freezing as a first option over drying. We would also not repeat the claim that it kills a starter, because the one controlled comparison we could find says the opposite.

How long dried sourdough starter really lasts

Dried starter is the best backup available, and the decade-long shelf life everyone quotes traces back to a single sentence that hedges itself twice.

What is actually established, ranked by support:

• 60 days, freeze-dried, measured. A 2024 study in Microorganisms found freeze-drying itself cost under 0.6 log CFU per gram, and that after 60 days of storage the LAB had fallen about 1 log and the yeasts 1 to 2 log.
• 6 months, freeze-dried, measured. Reale and colleagues recorded a decline of 3 log CFU per gram after six months at room temperature, from starting populations of 9.17 log for LAB and 7.53 for yeast.
• Over 2 years, air-dried, first-hand account. Maurizio Leo states he has kept a dried starter that long and revived it.
• "Over a decade," hearsay. King Arthur's own wording is that they are not sure, but they have heard cases of dried starter remaining viable for over a decade. That hedge gets dropped everywhere the number is repeated.

Notice the pattern in the measured numbers. The yeast loses more than the bacteria again.

To dry a starter, follow King Arthur's method. Spread ripe starter thin on parchment, let it dry at room temperature until it snaps between your fingers, then store the flakes airtight in glass somewhere cool and dark. Heat-drying is worse: oven drying at 40°C for 20 hours left yeast cells unable to restore full functionality, and spray drying with a 90°C outlet temperature cost 4 log CFU per gram.

Whether to pour off hooch or stir it back in

Pour it off if it is thick and dark, stir it in if it is thin and clear. That is the only operational rule any named authority publishes, and the sources otherwise contradict each other.

First, what hooch is. King Arthur describes it as mostly alcohol and water, with starch, sugar and some undigested flour suspended in it, and explains the color: the longer the liquid sits, the more those particles oxidize and the darker it gets. Darkness indicates elapsed time, not danger. We could find no peer-reviewed compositional analysis of starter hooch, so that description is the best that exists.

Now the disagreement. Across four of its own posts King Arthur variously says the choice is up to you, leans toward stirring it in unless there is more than about half an inch, instructs you to pour off any discolored hooch, and describes draining off most of the liquid in a six-month revival. The Perfect Loaf offers both options without preferring either. The Sourdough Journey is the only source giving a criterion: thin and clear gets stirred in, thick and dark grey and boozy-smelling gets poured off.

Use Cucuzza's rule. There is a reason for it that none of the sources spell out: hooch is ethanol plus accumulated acid, precisely the compounds suppressing the yeast. Pouring off a thick dark layer removes some of that suppression before you feed. Stirring in a thin clear layer preserves flavor and hydration at negligible cost.

When to actually throw a starter away

Mold ends a starter and nothing else on this page does. The rule is fuzzy or brightly colored, and it is not a judgment call.

Three authorities converge here. King Arthur's sourdough FAQ says to throw it out if it molds, if the odor is not the usual clean sour aroma, or if it develops a pink or orange color, and notes that an alcohol smell is fine. The Perfect Loaf lists pink, orange or green colors, white fuzzy spots, and darker areas with white on top. Cucuzza puts it most usefully: the only thing that can really kill your starter is mold.

Refrigeration does not protect against this. The 2025 Foods review states directly that 4°C does not protect sourdough from filamentous fungi contamination. A jar in the fridge for three months can still grow mold.

Now the honest gap. Every sourdough article on the internet seems to carry a kahm-yeast-versus-mold comparison table, and we could not find a single sourdough authority publishing visual criteria for kahm yeast on a starter. The one that mentions it at all tells you to go and research the difference. Those tables trace back to vegetable-fermentation blogs, not to bread sources.

So we are not publishing one. Fuzzy texture, or pink, orange, green or blue color, means discard the jar, wash everything and start again. For anything ambiguous on a culture you can replace in a week, discard. A new starter costs you flour and eleven days.

Sources

Related reading

• Sourdough starter smells like acetone? The specific smell a neglected starter develops, and why a bigger feed clears it.
• Why is my sourdough starter not rising? Where to go if the revival stalls at the rise stage.
• The best flour for a sourdough starter What to feed it during revival and afterward.

FAQ